Chromothripsis Challenges the Germline

Late Breaking Chromosomes

Chromothripsis Challenges the Germline

Complex chromosomal rearrangements (CCRs) constitute a rare, difficult to detect and hard to interpret entity in clinical cytogenetic diagnosis. CCRs are usually detected by inspection of chromosomal banding patterns during karyotyping. The sensitivity of detection is thus limited by the number of microscopically visible bands, typically in the order of 500 to 800. Array-CGH offers a dramatically improved resolution of detection but is limited to CCRs that are associated with copy number variations (CNVs), i.e. losses or gains. Copy neutral CCRs, in contrast, escape detection by array-CGH [Hochstenbach et al., 2009, 2011]. CNVs and CCRs are basically thought to result from nonallelic homologous recombination (NAHR) of unique DNA segments flanked by specific architectural features known as low-copy repeats, which are similar to the segmental duplications that arose during chromosomal evolution [Sharp et al., 2008; Liu et al., 2012]. This mechanism generates recurrent rearrangements, such as deletions, duplications, inversions, and translocations, with identical breakpoints, size and gene content because they are flanked by homologous low-copy repeats. All nonrecurrent genome rearrangements are, by exclusion, thought to be caused by replicative mechanisms, such as microhomology mediated break induced repair and nonhomologous end joining (NHEJ), which may involve long interspersed elements, retrotransposition, mobile elements, or the like [Liu et al., 2012]. Currently, investigators assume a dichotomy of NAHR-driven ‘simple’ rearrangements and replication-based CCRs. A recent study by Kloosterman et al. [in press] challenges such a simple dichotomy and suggests that CCRs should be viewed in their own right, rather than as an extension of reciprocal rearrangements and CNVs.

Kloosterman et al. [in press] have applied a novel technique, mate-pair sequencing, to the study of structural genome variation [Kloosterman et al., 2011]. This technique involves shearing the entire genomic DNA into pieces of a precisely defined size, for instance 3 kbp. Subsequently, a ‘genomic library’ is made and subjected to massively parallel sequencing. During this step, small stretches of DNA, typically 50 bp, on either end of every member of the library are being sequenced and mapped onto the Human Reference Genome. Pairs of ends which do not map to the preset distance of each other, are in the ‘wrong’ order, or on 2 different chromosomes reflect structural variations. For instance, pairs which map further apart than in the reference genome cover a genomic loss, while pairs mapping in the inverse order indicate an inversion [Medvedev et al., 2009]. After bioinformatic processing, the aberrantly mapping pairs can be identified, and with PCR, using the mapped ends to design primers, the breakpoints can be determined with nucleotide resolution. This technique had been successfully applied to the genome of a patient with a 3-way translocation. In addition to the expected 3 translocation breakpoints, Kloosterman et al. [2011] found 9 more aberrations. The breakpoints were consistent with a series of simultaneous double-stranded DNA breaks followed by fusion of the resulting chromosomal fragments by NHEJ. This phenomenon had previously been observed in tumors [Stephens et al., 2011] and was termed ‘chromothripsis’ (chromosome shattering).

Based on their studies of 11 patients in total, Kloosterman found a common signature for 9 cases [Kloosterman et al., 2011, and in press]. In those instances, the junctions of the CCRs contained no or only very short stretches of homology, were either copy neutral or contained only losses, and often involved an intricate network of breakpoints. The absence of flanking homologies and duplications sets these cases apart from the ‘classical’ NAHR-driven and replication-based mechanisms of CNV and CCR formation [Liu et al., 2011, 2012]. The finding of short-spaced intricate networks of breakpoints distinguishes these cases from replication-based mechanisms [Liu et al., 2012]. Such networks of breakpoints have earlier been observed but left investigators puzzled [Houge et al., 2003; Schluth-Bolard et al., 2009; Poot et al., 2010]. In particular, findings of multiple losses on a single or on several chromosomes in association with translocations or inversions left scientists at a loss for an explanation [Houge et al., 2003; Borg et al., 2005; Gribble et al., 2005; De Gregori et al., 2007; Poot et al., 2009, 2010]. In hindsight, these case reports and case series may indeed have described ‘chromothripsis’, a process being loosely defined as chromosome shattering followed by NHEJ of the pieces. In such a process, the orientation of the stretches of DNA at the junction points can either be head-to-head or head-to-tail or tail-to-head or tail-to-tail [Kloosterman et al., 2012]. In a replication-based mechanism only head-to-tail and tail-to-tail junctions can occur. In chromothripsis, in contrast, a few stretches of DNA may ‘drop-out’, i.e. become losses, during the NHEJ of the stretches of DNA. This model provides a molecular description of a specific type of chromosomal rearrangement, the CCRs, which is distinct from the existing explanations for the origin of CNVs and simple reciprocal translocations. Moreover, chromothripsis is a novel concept and not a mere extension of existing models such as NAHR and replication-based mechanisms.

Viewed as a novel mechanism of chromothripsis in its own right, chromosome shattering due to double-strand breaks followed by NHEJ of the resulting pieces presents us with both molecular and clinical challenges. Recurrent CNVs and CCRs are flanked by well-defined architectural features, while chromothripsis appears to occur at random, which makes its frequency in the population and the position of its breakpoints difficult to assess. Like CNVs, chromothripsis not only leads to losses of genes, but may also disrupt genes or may, by altering the position of a gene by inversion or translocation of a chromosomal fragment, alter the control of its expression [Poot et al., 2011]. Therefore, the phenotypic outcome of chromothripsis may be much more difficult to predict than that of NAHR.

During the past decade or so, molecular cytogenetics have focused on CNVs or inversions and translocations as separate entities, such that the field remained oblivious for the effects of combinations of several types of structural rearrangements in a single case. Only a few studies [Borg et al., 2005; Gribble et al., 2005; Poot et al., 2009, 2010; Granot-Hershkovitz et al., 2011] assessed the full complexity of structural variations by combining several molecular cytogenetic techniques, albeit in limited numbers of cases and at limited resolution. Mate-pair sequencing is currently the only technique able to asses all types of structural variation simultaneously and at the nucleotide level of resolution. In this way, the studies by Kloosterman et al. [2011, 2012] may prove paradigmatic for future molecular cytogenetic investigations, aiming to face the challenges presented by chromothripsis.